在CNN(Keras)中选择步幅和过滤器的数量 [英] Selecting number of strides and filters in CNN (Keras)

问题描述

我正在使用keras构建用于信号分类的cnn模型.在keras中进行超参数调整和选择步幅和数字过滤器的最佳方法是什么.

I am using keras to build a cnn model for signal classification. What is the best way in keras for hyper parameter tuning and selection for the number of strides,and number filters.

推荐答案

欢迎来到深度学习的主要问题.没有适合所有问题的有效单一解决方案.不过有一些模式，例如从早期的几个过滤器开始，增加过滤器数量，同时减小尺寸. 对于您来说，最好的方法是开始阅读现有的体系结构，例如Inception，VGG，Resnet等，以了解它们为何以及如何实现其网络.然后，通过遵循/发明自己的网络，您可以尝试获得更好的结果.

Welcome to main question of deep learning. There is no valid, single solution which fits to all problems. There are some patterns though, like starting with few filters in early layers and increase filter count while reducing the sizes. For you, the best would be to start reading existing architectures like Inception, VGG, Resnet, etc. to understand why and how they implemented their network. Then, by following / inventing your own network, you can try for better results.

这篇关于在CNN(Keras)中选择步幅和过滤器的数量的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！